Developer releases Sigil, a steganography tool that embeds cryptographic ownership signatures into image pixels to prevent unauthorized AI training data scraping

Sigil is a local-first desktop app built with Svelte, Tauri, and SQLite that hides ownership IDs in the least significant bit (LSB) layer of images using HMAC-SHA256 signing

The desktop vault remains closed-source to protect the signing architecture, but the creator open-sourced the Rust extraction standard for AI data procurement teams

The extraction tool allows organizations to scan scraped datasets and detect cryptographically locked assets that require licenses, using memory-safe pixel parsing via Rust image and hex crates

The project proposes an open protocol where data scrapers can identify protected digital assets before using them for training AI models